The Foundation
Core concepts that underpin every conversation you'll have at Point Predictive. Get these right and everything else makes sense.
Company Overview
Who Point Predictive is, where the company came from, and why it exists. Understand this before anything else.
Point Predictive is a leading provider of artificial intelligence-driven fraud prevention, risk intelligence, and verification solutions focused primarily on the consumer lending and automotive finance industries. The company helps banks, credit unions, captives, fintechs, and specialty lenders reduce fraud losses, improve underwriting efficiency, and safely grow loan originations.
What differentiates Point Predictive in the market is its consortium-based intelligence network and real-time application visibility across lenders, dealers, and finance sources. By analyzing patterns across a broad ecosystem of lending activity, Point Predictive can identify fraud risks and emerging attack patterns that individual institutions or traditional tools often cannot detect on their own.
The company's solutions are designed not only to stop fraud, but also to enable "safe growth" — helping lenders approve more legitimate borrowers while reducing losses tied to synthetic identity fraud, straw borrowers, credit washing-enabled bust-out fraud, dealer manipulation, and organized fraud rings.
Point Predictive was founded to address a growing problem in consumer lending: traditional fraud and credit risk tools were no longer sufficient to detect sophisticated, organized fraud attacks occurring across multiple lenders and channels.
As digital lending expanded and indirect auto finance became increasingly automated, fraudsters began exploiting gaps between lenders, dealers, and verification systems. Existing tools largely relied on static or backward-looking data such as credit bureau files, identity databases, and institution-specific historical performance. These approaches often failed to identify coordinated fraud behavior occurring in real time across the broader market.
Point Predictive was created to solve this challenge through advanced machine learning, artificial intelligence, and consortium data sharing. By aggregating and analyzing application activity across a large network of lenders and dealerships, the company developed the ability to detect hidden fraud patterns, suspicious application velocity, identity inconsistencies, and organized fraud networks much earlier in the lending lifecycle.
Over time, the company expanded its capabilities beyond application fraud detection into areas such as:
- →Income and employment validation
- →Dealer risk intelligence
- →Synthetic identity detection
- →Bust-out fraud identification
- →Fraud consortium analytics
- →Real-time risk scoring and alerting
- →Safe Loan Origination and Growth
Today, Point Predictive supports hundreds of financial institutions across the United States, including banks, credit unions, captive finance companies, fintech lenders, and automotive lenders.
Point Predictive's mission is to help lenders make safer, smarter, and faster lending decisions through the power of artificial intelligence, consortium intelligence, and real-time risk visibility.
The company aims to:
- →Reduce fraud and credit losses
- →Increase safe loan approvals
- →Improve operational efficiency
- →Help lenders identify risk earlier in the lifecycle
- →Provide visibility beyond a lender's own internal data
A core philosophy behind Point Predictive's approach is that fraud today is highly coordinated and increasingly cross-institutional. Because individual lenders can only see their own applications and customer behavior, many sophisticated fraud patterns remain invisible when viewed in isolation.
Point Predictive's consortium model helps solve this problem by enabling broader market intelligence and pattern recognition across lenders and dealers in real time.
"The company's broader vision is to transform lending from reactive fraud detection to proactive risk intelligence — helping financial institutions safely grow while protecting both lenders and consumers from increasingly sophisticated fraud threats."
These three values define how Point Predictive operates and what's expected of everyone on the team. They're not aspirational — they're the standard.
This doesn't mean you need to be an expert today, but we're looking for hungry people who constantly want to learn about our data, our products, and our market. This isn't a place where your hand will be held, and you're expected to take initiative.
Be an owner, even if it's not your job to do so. At Point Predictive you'll most likely be pulled into Sales, Marketing, Customer Success, and maybe even a little Data Science. We don't shy away from uncomfortable and like to get our hands dirty. Always treat deliverables with urgency and don't procrastinate. Startups will eat people alive who wait until the last minute.
Frank and Tim like to say "Get it done, done." We want to think about a project being complete from the eyes of the customer or end user, not at the end of our job description. We are a team here, and so everyone is willing to go the extra mile to ensure a project is complete and has a standard of excellence. Don't be afraid of confrontation and hold others to a higher standard of urgency and quality.
The Lending Landscape
A crash course in how lending actually works — understand the ecosystem before you sell into it.
Lender → Borrower
The borrower applies directly to the lender — walking into a bank or credit union branch, applying online, or calling in. The lender owns the full relationship from application through payoff.
Lender → Dealer → Borrower
The borrower applies at a dealership. The dealer submits the application to one or more lenders on the borrower's behalf. The lender never meets the borrower — this is the dominant model in auto finance.
Most of Point Predictive's market operates in indirect auto lending. The dealer sits between the lender and borrower, which creates a visibility gap — and a fraud opportunity. Dealers can manipulate applications in ways a direct lender would catch far more easily.
These are the numbers your prospects live and die by. Learn them — they're your entry points into every discovery conversation.
"Where are you seeing unexpected losses?" is often the best first question. Charge-off spikes and delinquency trends are trigger events — they create urgency in a way that cold outreach rarely does.
Auto lending involves four distinct groups. Know who each one is and how they interact — your prospects fall into the lender category, but you'll often hear about the others in every conversation.
Banks, credit unions, captive finance companies (e.g., Toyota Financial), fintechs, and specialty lenders. They fund loans, take on credit risk, and absorb fraud losses. They are your buyers.
Automotive dealerships originate the majority of auto loan applications. They submit applications to lenders through platforms like DealerTrack and RouteOne. Some dealers are bad actors — falsifying income, inflating values, or colluding with borrowers.
The individual applying for the loan. Most are legitimate — but a meaningful percentage misrepresent their income, employment, or identity. First-party fraud (the borrower lying) is far more common than most lenders realize.
Loan Origination Systems are the software lenders use to receive, process, and decision applications. MeridianLink, Sync1, DeFi Solutions, and Origence are common examples. Point Predictive integrates directly into these platforms — knowing a prospect's LOS is a key qualification question.
Understanding the loan lifecycle tells you where fraud enters and when it shows up as a loss. These are two very different moments.
"Fraud occurs at application — but losses show up months later. By the time a lender sees the loss, the fraudster is long gone."
The core timing problem Point Predictive solvesThis timing gap is why most lenders underestimate their fraud exposure. They see the loss in their charge-off report and call it a credit problem. They never connect it back to a fraudulent application that slipped through months earlier.
Fraud losses are routinely miscategorized as credit losses because most lenders never identify the fraud that caused the default. This means the true cost of fraud is almost always understated — and lenders are often spending money solving the wrong problem.
Every loan decision involves some combination of automated and manual underwriting. Understanding this helps you identify where inefficiency — and fraud exposure — lives in a prospect's process.
Rules + Scores
Credit scores (FICO), bureau data, debt-to-income ratios, and policy rules run automatically inside the LOS. Fast and consistent, but only as good as the data it's fed — and easily gamed by borrowers who know the thresholds.
Human Review
An underwriter manually reviews income documents, pay stubs, and employment records for loans that don't auto-approve. Time-consuming, inconsistent, and still vulnerable to fabricated documents. IEValidate automates this layer.
Most lenders already have fraud and risk tools in place. The problem isn't that they have nothing — it's that what they have was built for a different era of fraud.
Measures creditworthiness — the likelihood a borrower will repay based on past behavior. It says nothing about whether the application information is truthful or whether the borrower is part of a fraud network.
Verifies that the person applying is who they say they are. Excellent at catching third-party identity theft — but completely blind to first-party fraud, where the real person applies and simply lies about their income or employer.
Built on a single lender's historical data — which means they can only detect fraud patterns that have already occurred at that institution. Organized fraud rings specifically target multiple lenders simultaneously to stay under each one's radar.
"Any tool that only sees your data can only find your fraud. The patterns that span multiple lenders are invisible to everyone — except a consortium."
The gap Point Predictive fillsFraud in auto lending isn't a single problem — it's a collection of distinct attack types, each requiring different detection approaches. Know these before your first discovery call.
A fraudster builds a fake identity using a mix of real and fabricated information — often a real SSN (frequently a child's) combined with a fabricated name and address. The synthetic identity builds credit over time before being used to take out loans with no intention of repayment.
A person with good credit applies for a loan on behalf of someone who couldn't qualify — often for compensation. The straw borrower has no intention of making payments. This is particularly common in dealer-assisted fraud schemes.
Fraudsters dispute accurate negative items on their credit report (a process called "credit washing") to artificially inflate their score, then apply for multiple loans across multiple lenders simultaneously and default on all of them. Point Predictive can identify this pattern before funding.
Dealers manipulate applications by falsifying income documents, inflating vehicle values, fabricating employment, or colluding with borrowers. Some dealers operate as organized fraud rings, routing fraudulent applications across multiple lenders. DealerCheck is designed to surface these patterns.
Every one of these fraud types is most effectively detected using cross-lender, real-time data. A single lender's internal view catches some of it — but coordinated attacks are designed specifically to stay under any one lender's threshold. That's the problem Point Predictive was built to solve.
What Makes Point Predictive Different
Four things no competitor can replicate. Know these before your first prospect call.
For years, the lending industry focused almost exclusively on identity fraud -- the fraud they could see, tag, and report. Point Predictive went a different direction. We focused on early payment default as the primary signal, not confirmed fraud tags. That was against the grain at the time.
Most fraud never gets tagged as fraud. A borrower lies about their income, the loan funds, and three months later it defaults. The lender calls it a credit loss and moves on. The fraud is invisible -- but EPD is the footprint it leaves behind.
"Most fraud tools wait for lenders to identify fraud and then learn from those tags. We built our models around early payment default -- because that's where the fraud actually shows up, whether or not anyone called it fraud."
Point Predictive has spent a decade building the largest lending fraud consortium in the country. 650+ financial institutions contribute application data every month, creating a pool of intelligence that now covers 307 million scored applications and over 90 billion unique data points.
This isn't something a competitor can replicate by hiring engineers or raising capital. It took years of trust-building with lenders, years of data accumulation, and years of fraud tagging to get here. The consortium is the moat. No other vendor is close.
"Our competitors can build a model. They can't build a decade of consortium data. That's not a technology gap -- it's a time gap. And it only grows."
Most competitors build custom machine learning models for each lender. These models are overfit to that lender's historical data, need constant maintenance, and struggle to adapt when fraud patterns shift.
Point Predictive uses a deep learning LSTM (Long Short-Term Memory) approach with common models trained across the entire consortium. Every lender benefits from intelligence that spans the whole network, not just their own history.
Beyond the models, Point Predictive has built an Army of Fraud Agents -- AI agents each trained to identify a very specific type of fraud. These agents comb through the millions of applications processed each month, surfacing emerging trends in a fraction of the time it would take a human team. They operate 24/7.
"A custom model trained on your data is only as good as your fraud history. Our common model is trained on the entire industry's fraud history. There's no comparison."
Point Predictive was built by fraud experts, not data scientists who later added fraud as a feature. The company has a team of seasoned fraud investigators who review applications daily -- and they operate as an extension of every customer's fraud team.
When a new fraud trend is identified, it gets fed directly to the data science team so the trend can be captured in the models in real time. The human expertise and the technology operate as a closed loop -- each making the other sharper.
"When you work with Point Predictive, you're not just buying a score. You're getting a fraud team that monitors the entire lending ecosystem on your behalf, every single day."
"Any vendor can sell you a model. Point Predictive brings the data, the technology, the expertise, and the approach that was built to find fraud that most lenders don't even know they have."
How to tie all four togetherFirst-Party vs. Third-Party Fraud
The most important distinction in this job -- and the one new sellers get wrong most often.
The Impersonator
Someone steals another person's identity to take out a loan. The victim has no idea. The fraudster is a criminal pretending to be someone else. This is what most people picture when they hear "fraud."
The Liar
The real person applies -- but lies on the application. Identity checks out. But they fabricated income, used a fake employer, or never intended to repay. It funds, then it defaults.
The industry spends almost all of its time and money solving third-party fraud. LexisNexis, SentiLink, Experian -- excellent identity tools, table stakes at this point. But according to Point Predictive's 2026 data, 69% of auto lending fraud exposure is first-party fraud. Income and employment misrepresentation alone accounts for 45% of total exposure -- $4.68B annually.
Most lenders don't know this. They think they have a fraud problem if someone steals a customer's identity. They don't realize the bigger problem is sitting in their credit loss bucket.
"Most fraud tools verify who someone is. Point Predictive verifies what they're telling you is true."
The line you'll use constantlyMost lenders focus on the fraud they can see and catch. First-party fraud is invisible until a loan defaults -- and by then it's been miscategorized as a credit loss. Your job is to show them what they can't see.
The Consortium
The single biggest reason Point Predictive's models outperform anything a lender could build themselves.
What It Is
650+ financial institutions -- banks, credit unions, auto lenders, fintechs -- all submitting application data to a shared pool. Point Predictive analyzes that data, identifies fraud patterns across all of it, and feeds those signals back into the models every lender uses when they run a score.
Why It Matters
Fraud is invisible in isolation. A lender might see one application from a borrower -- no red flags on its own. But the consortium sees all 8 applications that same borrower submitted across 4 different dealerships in the same week. That cross-lender pattern becomes a high-confidence fraud signal. No single institution can build that alone.
"Fraud rings count on lenders not talking to each other. Our consortium is how lenders talk to each other without sharing anything they shouldn't."
How to explain consortium value simplyWhat's Shared -- and What Isn't
Point Predictive is SOC 2 Type II certified and has never had a security incident. Security is the #1 deal blocker -- get ahead of it early.
Terminology
Terms that will come up constantly -- in prospect conversations, internal discussions, and demos. Click any term to expand.
The total number of loan requests a lender receives -- the top of the funnel. Every downstream metric flows from this number, so it's usually the first thing you establish when sizing a prospect's business.
The number of applications that pass a lender's credit policy and receive a credit decision. Approvals don't mean funded loans -- they just mean the lender said yes.
The number of approved loans that actually close and get funded. This is the number that drives revenue. The gap between approvals and funded loans is where friction, stipulations, and competitive loss all show up.
Approvals divided by total applications. A first look at how conservative or aggressive a lender's credit policy is.
Funded loans divided by approvals. Tells you how well a lender converts approved deals into actual business. A low capture rate usually means deals are being lost before funding -- most often due to friction, slow response times, or stipulation overload.
"If you're approving 1,000 applications a month but only funding 400, that's 600 deals you said yes to and still lost. How much of that is stipulations slowing you down?"
Funded loans divided by total applications. The most complete picture of how effective a lender is overall -- from application all the way through to a closed loan.
When applications come to a lender through a dealer rather than directly from the borrower. The dealer generates the application and shops it to multiple lenders simultaneously -- the lender originates the loan once it's funded. This creates intense competition. The lender who responds fastest with the least friction usually wins the deal.
"In indirect lending, you don't just compete on rate. You compete on speed and simplicity. A dealer will take a slightly worse rate from a lender who doesn't bury them in stip requests."
When a lender consistently receives the deals that everyone else already passed on. Dealers prioritize their primary lending partners -- those lenders get first-look at the best applications. Secondary and tertiary lenders see what's left.
"If your EPD rate is significantly higher than the consortium average for similar dealers, that's often adverse selection at work."
A loan that stops making payments within the first 3-6 months of origination. Strongly correlated with application fraud. 70% of EPDs contain evidence of fraud or misrepresentation -- but lenders categorize them as credit losses, making the fraud invisible.
"What's your current early payment default rate? And how are you currently categorizing those -- as fraud or credit losses?"
When someone applies for a loan on behalf of another person who can't qualify themselves. The real borrower intends to use the vehicle; the straw borrower is just the name on the application.
"A lender receives a single borrower application -- no obvious red flags. But Point Predictive saw that same borrower apply 30 minutes ago with a co-borrower at a different dealer and get declined. The co-borrower didn't disappear -- they're the one the car is actually for. That's a straw purchase."
A fabricated identity built by combining a real SSN with a false name and address. The identity gets established slowly, builds credit, then gets used to take out loans with no intent to repay.
An organized scheme where a borrower or fraud ring builds credit relationships with multiple lenders simultaneously, then maxes everything out at once and disappears. Up 83% since 2021. Cross-lender by nature -- almost impossible to catch without consortium data.
When a borrower disputes accurate negative information on their credit report to artificially improve their score before applying for a loan. The disputes are fraudulent but the process appears legal, which makes it hard to catch.
When a dealer inflates the listed price or features of a vehicle to secure a larger loan than the car warrants. Dealers with high powerbooking activity have EPD rates of 8% vs. 2% for clean dealers.
Conditions a lender places on a loan before funding it. A high stipulation rate slows down the loan process and creates friction. Point Predictive reduces stips by using consortium intelligence to automatically clear applicants who don't need manual review.
The percentage of applications that score positively in a verification system. For IEValidate, it refers to what percentage of applicants can be verified without additional documentation. Higher hit rate = less friction = better borrower experience.
Equifax's employment and income verification database. Primary competitor to IEValidate. Covers roughly 30-40% of borrowers. Charges $10-$20+ per verification. Has no fraud detection capability.
The Fair Credit Reporting Act and Gramm-Leach-Bliley Act -- the two primary regulatory frameworks governing how consumer financial data can be used. Point Predictive's FCRA-compliant model allows scores to be used in adverse action decisions. Many competitors' models aren't FCRA-compliant.